Dear netbsd/hpcarm hackers,

I stopped hacking my jornada 820 for two weeks but now I'm back to it.
(Well, I have used it as a mobile Emacs/WinCE console, but nothing more.)

It seems that hpcboot cannot directly run linux (or rather,
that linux cannot directly start after being hpcboot-loaded),
so a linux kernel hacker told me to port blob
as an intermediate loader between hpcboot and linux
-- which also makes for a simpler and more manageable project.

Whereas loading vmlinux.gz would clobber the framebuffer with the
(300 first KB of the) linux kernel then die, loading blob-chain-elf32.gz
kind of "works" before it dies: I could get blob to output a few characters
on the serial console before it would die; the faster the serial line,
the more characters had the time to get through.
(I seem to not be able to reproduce that effect anymore).
Erik Mouw, the blob author suggests that hpcboot might not successfully
enter SVC mode. The fact that hpcboot reports CPSR :600000df is suspect:
the last bits for SVC mode are 10011, and the reported bits 11111 are
not defined in my ARMv7.pdf spec sheet.
In any case, instead of clobbering the framebuffer, booting blob
switches the LCD off (maybe an artefact of blob).

Can you suggest things to try to debug hpcboot, blob and/or linux?
What about cross-compiling NetBSD/hpcarm from Linux?

As usual, my pathetic attempts are available at
	ftp://Samaris.tunes.org/jornada820/

I welcome any information, advice, tidbits, etc.

[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
[  TUNES project for a Free Reflective Computing System  | http://tunes.org  ]
The meta-Turing test counts a thing as intelligent if it seeks to apply
Turing tests to objects of its own creation.
		-- Lew Mammel, Jr.